Introduction

Polythiophenes have been studied as functional materials in electronic devices, including organic field-effect transistors (OFETs), organic light-emitting diodes and organic photovoltaics.1, 2, 3, 4 Of the known polythiophenes, regio-regular poly(3-hexylthiophene-2,5-diyl) (P3HT in Scheme 1) is of particular importance. The high planarity of its polymer backbone leads to the extended π-conjugation and enhances the interchain interaction in the solid state. In addition, the high solubility of P3HT provides excellent processability, making it suitable for use in organic electronic devices.1, 3

On the other hand, heteroatom-bridged 2,2′-bithiophene derivatives, such as dithienophosphole,5 -borole6 and -pyrrole,7 have been extensively studied as building units of conjugated oligomers and polymers. In these systems, the bithiophene unit is fixed to be highly planar by the bridge to enhance the conjugation. In addition, the bridging element often exhibits interesting electronic effects on the bithiophene electronic state. Previously, we introduced silicon- and germanium-bridged bithiophenes (dithienosilole (DTS) and dithienogermole (DTG) in Scheme 1),8, 9 in which the bonding interaction between Si or Ge σ* and bithiophene π*-orbital (σ*–π* conjugation) stabilizes LUMO (lowest unoccupied molecular orbital), whereas HOMO (highest occupied molecular orbital) is also lowered, although only slightly. As a result, DTS and DTG show smaller HOMO–LUMO energy gaps than their carbon analog (cyclopentadithiophene (DTC) in Scheme 1). Based on these electronic characteristics, DTS and DTG have been used as building units of functional π-conjugated polymers. In particular, the use of donor–acceptor polymers with DTS and DTG as the donor, as the active layer of bulk heterojunction polymer solar cells and OFETs, has been explored.9, 10, 11, 12 DTS homopolymers were also studied independently by our group13 and Marks and colleagues14 (Scheme 1), and were found to have a more expanded conjugation (λmax=533 nm for pDTS3 in Scheme 1) than P3HT (λmax=458 nm).14 Recently, we prepared DTG homopolymers for the first time and applied them to bulk heterojunction polymer solar cells and OFETs (pDTG1 and pDTG2, Scheme 1).15 OFET devices made of the DTG homopolymers showed clear p-type transistor response. This contrasted the previous finding that pDTS3 (R=n-hexyl) exhibited no transistor response in spite of its extended conjugation.14 It seemed that the nature of the bridging elements affected the semiconducting properties of the films, although it was also likely due to the lower molecular weight of pDTS3 (Mn=9000) than that of pDTG2 (Mn=79 000) and the shorter side chains of the bridging silicon. In the present study, we prepared new DTS homopolymers with the same alkyl chains (pDTS1 and pDTS2 in Scheme 1) as and similar molecular weights to those of DTG homopolymers to enable direct comparison of DTS and DTG polymers. We also prepared a DTC homopolymer with n-octyl groups for comparison (pDTC2 in Scheme 1).

Experimental procedure

General

All reactions were carried out in dry argon. Chlorobenzene (purity 99.5%), toluene (99.0%) and ether (99%) that were used as the reaction solvents were purchased from Kanto Chemical Co. Inc. (Tokyo, Japan) and Wako Pure Chemical Industries, Ltd (Osaka, Japan), respectively. They were distilled from calcium hydride and stored over activated molecular sieves until use. Starting compounds 2,6-dibromo-4,4-bis(2-ethylhexyl)dithienosilole (DTS1Br2),10 2,6-dibromo-4,4-dioctyldithienosilole (DTS2Br2),16 4,4-dioctyl-2,6-bis(trimethylstannyl)dithienosilole (DTS2Sn2),17 and 2,6-dibromo-4,4-dioctyl-4H-cyclopenta[1,2-b:5,4-b']dithiophene (DTCBr2)18, 19 were prepared as reported in the literature. To remove palladium species, the filtered reaction mixtures were heated with a solution of sodium N,N-diethyldithiocarbamate trihydride (3.1 g, Kanto Chemical Co. Inc., purity 92%) in 30 ml of water at 80 °C with stirring for 2 h. Then, the organic layer was separated and washed in the following order: with water, 3% acetic acid aqueous solution and water again. The organic layer was dried over anhydrous magnesium sulfate and the solvent was removed under vacuum. Nuclear magnetic resonance (NMR) spectra were recorded on Varian 400-MR and System-500 spectrometers (Agilent Inc., Santa Clara, CA, USA). Atmospheric pressure chemical ionization (APCI) mass spectra were measured on a Thermo Fisher Scientific LTQ (Waltham, MA, USA) Orbitrap XL spectrometer at the Natural Science Center for Basic Research and Development (N-BARD) Hiroshima University (Higashi-Hiroshima, Japan), and electron ionization mass spectra were recorded on a Shimadzu QP-2020 A spectrometer (Shimadzu Co., Kyoto, Japan). Ultraviolet–visible (UV–vis) absorption and photoluminescence (PL) spectra were measured on Hitachi U-3210 and Hitachi F-4500 spectrophotometers, (Hitachi, Ltd., Tokyo, Japan)respectively. Polymer molecular weights were determined by gel permeation chromatography (GPC) using serially connected Shodex (Showa Denko K.K., Tokyo, Japan) KF2001 and KF2002 columns and tetrahydrofuran (THF) as the eluent. OFET devices using the present polymers were fabricated as reported in the literature.20

Preparation of 6-bromo-4,4-bis(2-ethylhexyl)-2-(trimethylstannyl)dithienosilole (DTS1BrSn)

To a solution of DTS1Br2 (1.788 g, 3.101 mmol) in 35 ml of ether was added n-butyllithium solution (1.9 ml, 1.64 M in hexane, Kanto Chemical Co. Inc.) slowly at −80 °C. After stirring the mixture at this temperature for 30 min, trimethyltin chloride (0.718 g, 3.601 mmol, Tokyo Chemical Industry Co., Ltd (Tokyo, Japan), purity >98%) was added and the mixture was allowed to warm to room temperature. The mixture was poured into ice water, and the organic layer was separated and dried over anhydrous magnesium sulfate. After the solvent was evaporated, the residue was purified by preparative GPC with toluene as eluent to give 1.346 g of a diastereomeric mixture of DTS1BrSn as a yellow viscous oil (66% yield); 1H NMR (δ in CDCl3, 500 MHz) 0.37 (s, 9H), 0.77 (t, J=7.0 Hz, 6H), 0.79–0.86 (m, 6H), 0.86–0.97 (m, 4H), 1.08–1.30 (m, 16H), 1.36–1.42 (m, 2H), 6.97 (s, 1H), 7.07 (s, 1H); 13C NMR (δ in CDCl3, 125 MHz) −8.18, 10.80, 14.15, 17.50, 17.54, 17.62, 17.65, 22.94, 22.99, 28.76, 28.85, 28.87, 28.89, 35.52, 35.57, 35.77, 35.85, 110.37, 110.39, 110.41, 132.37, 132.39, 132.40, 137.65, 137.68, 137.70, 138.37, 138.39, 138.40, 142.43, 142.57, 142.73, 143.05, 143.09, 143.14, 149.45, 149.50, 153.93, 153.94 (coupling with Sn is included); high-resolution mass spectrometry (APCI) Calcd: [M+] 660.09318. Found: 660.09192.

Preparation of 6-bromo-4,4-dioctyl-2-trimethylstannanyl-4H-cyclopenta[1,2-b:5,4-b']dithiophene (DTC2BrSn)

To a solution of DTC2Br2 (0.561 g, 1.001 mmol) in 12 ml of ether was added n-butyllithium solution (0.65 ml, 1.55 M in hexane) slowly at −80 °C. After stirring the mixture at this temperature for 30 min, trimethyltin chloride (0.219 g, 1.100 mmol) was added and the mixture was allowed to warm to room temperature. The mixture was poured into ice water, and the organic layer was separated and dried over anhydrous magnesium sulfate. After the solvent was evaporated, the residue was purified by preparative GPC using toluene as eluent to give 0.387 g of DTC2BrSn as a yellow viscous oil (60% yield); 1H NMR (δ in CDCl3, 400 MHz) 0.38 (s, 9H), 0.85 (t, J=7.2 Hz, 6H), 0.93–1.00 (m, 4H), 1.15–1.26 (m, 20H), 1.76 (t, J=8.2 Hz, 4H) 6.92 (s, 1H), 6.93 (s, 1H); 13C NMR (δ in CDCl3, 100 MHz) −8.02, 14.12, 22.66, 24.52, 29.27, 29.31, 29.98, 31.82, 37.52, 53.53, 109.98, 124.85, 129.06, 137.07, 138.28, 141.73, 156.64, 159.59 (coupling with Sn is included); high-resolution mass spectrometry (APCI) Calcd: [M+] 644.11625. Found: 644.11523.

Preparation of pDTC2

A mixture of DTC2BrSn (0.202 g, 0.313 mmol), tris-(dizenzylideneacetone)dipalladium(0) (Pd2(dba)3, 14.3 mg, 0.016 mmol, Tokyo Chemical Industry Co., Ltd, purity >75%), tri(o-tolyl)phosphine (P(o-tol)3, 19.1 mg, 0.063 mmol, Wako Pure Chemical Industries, Ltd, >96%) and 10 ml of chlorobenzene was vacuum degassed and heated to reflux for 4 days. The mixture was cooled to room temperature and the resulting precipitates were filtered. After removal of the palladium species as described above, reprecipitation of the residue from chlorobenzene/methanol, chlorobenzene/ethanol and then chlorobenzene/acetone provided 45.1 mg (36.1% yield) of pDTC2 as black solids: m.p. >300 °C; 1H NMR (δ in C6D5Cl, 400 MHz) 0.86 (t, J=7.2 Hz, 6H, octyl), 1.09–1.21 (br m, 24H, octyl), 1.94 (br s, 4H, octyl), 7.25 (br s, 2H, thiophene); 13C NMR (δ in C6D5Cl, 100 MHz) 14.34, 23.01, 24.99, 29.73, 30.46, 32.16, 54.52 (sp2 carbons could not be observed because of low solubility); GPC Mn=12 000, Mw=28 000, Mw/Mn=2.3; UV–vis abs λmax=592 nm (in C6H5Cl).

Preparation of pDTS1

A mixture of DTS1BrSn (0.658 g, 0.996 mmol), Pd2(dba)3 (18.2 mg, 0.020 mmol), P(o-tol)3 (30.2 mg, 0.099 mmol), CuO (79.2 mg, 0.997 mmol, Wako Pure Chemical Industries, Ltd, purity 99.9%) and 20 ml of chlorobenzene was vacuum degassed and heated to reflux for 5 days. The mixture was cooled to room temperature and the resulting precipitates were filtered. After removal of the palladium species as described above, reprecipitation of the residue from chlorobenzene/methanol gave polymeric substances that were then placed in a Soxhlet apparatus and washed with hot methanol, ethyl acetate and hexane. Finally, the residue that remained insoluble in those hot solvents was extracted with hot benzene. The benzene solution was poured into ethyl acetate and the resulting precipitates were collected to give pDTS1 (0.337 g, 81%) as black solids: m.p. >300 °C; 1H NMR (δ in CDCl3, 500 MHz) 0.78–0.89 (m, 12H, 2-ethylhexyl), 0.94–1.05 (br m, 4H, 2-ethylhexyl), 1.13–1.35 (m, 16H, 2-ethylhexyl), 1.43–1.51 (br m, 2H, 2-ethylhexyl), 7.10 (br s, 2H, thiophene); 13C NMR (δ in CDCl3, 125 MHz) 10.82, 14.19, 17.61, 23.02, 28.88, 28.92, 35.63, 35.92 (sp2 carbon signals could not be observed because of low polymer solubility); GPC Mn=42 000, Mw=76 000, Mw/Mn=1.8; UV–vis abs λmax=569, 597 nm (in C6H5Cl).

Preparation of pDTS2

A mixture of DTS2Sn2 (0.651 g, 0.875 mmol), DTS2Br2 (0.504 g, 0.874 mmol), Pd2(dba)3 (16.0 mg, 0.017 mmol), P(o-tol)3 (26.0 mg, 0.085 mmol), CuO (69.6 mg, 0.875 mmol) and 20 ml of chlorobenzene was vacuum degassed and heated to reflux for 5 days. The mixture was cooled to room temperature and the resulting precipitates were filtered. After removal of the palladium species as described above, reprecipitation of the residue from chlorobenzene/methanol gave polymeric substances that were then placed in a Soxhlet apparatus and washed with hot methanol, ethyl acetate and hexane. Finally, the residue that remained insoluble in those hot solvents was extracted with hot benzene. The benzene solution was poured into hexane/benzene (3:2) and the resulting precipitates were collected to give pDTS2 (0.483 g, 66%) as black solids: m.p. >300 °C; 1H NMR (δ in C6D4Cl2, 500 MHz) 0.86 (t, J=5.0 Hz, 6H), 1.09–1.25 (br m, 24H), 1.52 (br s, 4H), 7.35 (br s, 2H); 13C NMR (δ in C6D4Cl2, 125 MHz) 14.35, 23.02, 24.56, 29.50, 29.66, 32.17 (sp2 carbon signals could not be observed because of low polymer solubility); GPC Mn=70 000, Mw=175 000, Mw/Mn=2.5; UV–vis abs λmax=603 nm (in C6H4Cl2).

Results and discussion

Polymer synthesis

Poly(dithienosilole-2,6-diyl)s pDTS1 and pDTS2, which have branched and linear solubilizing alkyl groups on the silicon atoms, respectively, were prepared, as shown in Scheme 2, similarly to what was previously reported for pDTG1 and pDTG2. The progress of polymerization was monitored by GPC and the reaction was quenched when the molecular weights no longer increased. All the polymers were obtained as black solids. Stille homocoupling of DTS1BrSn that was prepared by monolithiation of dibromide DTS1Br2 followed by stannylation of the resulting bromolithiodithienosilole with trimethyltin chloride in refluxing chlorobenzene using Pd2(dba)3/P(o-tol)3/CuO as the catalyst (Scheme 2) gave pDTS1. On the contrary, selective monolithiation of DTS2Br2 failed, similar to the DTG analog. Attempted monolithiation of DTS2Br2/DTG2Br2 always proceeded nonselectively to give mixtures of mono- and dilithiated compounds and the starting DTS2Br2/DTG2Br2. Therefore, pDTS2 was obtained by the cross-coupling reaction of DTS2Sn2 and DTS2Br2. For pDTG1 and pDTG2, fractional extraction using a Soxhlet apparatus provided higher and lower molecular weight fractions, pDTG1H and pDTG1L, and pDTSGH and pDTG2L, respectively, as summarized in Table 1.15 However, similar Soxhlet extraction separated only small amounts of low molecular weight fractions of pDTS1 and pDTS2. It was found that the introduction of branched 2-ethylhexyl substituents in place of n-octyl groups tended to decrease the molecular weights for both the DTS- and DTG-containing polymers. This is likely because the large 2-ethylhexyl substituents cover the reactive center of DTS and DTG to suppress the reactivity.

Table 1 Properties of present and reported polymers
Full size table

For pDTC2, monolithiation of DTC2Br2 followed by treatment of the resulting monolithiated DTC with trimethyltin chloride gave bromo(trimethylstannyl)cyclopentadithiophene DTC2BrSn cleanly, and this was polymerized by a similar palladium-catalyzed Stille coupling reaction in chlorobenzene using Pd2(dba)3/P(o-tol)3 as the catalyst (Scheme 2). The behavior of DTC2Br2 was different from that of DTS and DTG derivatives with n-octyl groups (DTS2Br2, DTG2Br2) whose monolithiation did not proceed selectively as mentioned above.

The solubility of the present polymers was dependent on their molecular weights and alkyl chains. pDTS2 bearing n-octyl groups with high molecular weight showed low solubility. It was slightly soluble in o-dichlorobenzene and nearly insoluble in chloroform, toluene, acetone, hexane, ethyl acetate and ethanol at room temperature, but soluble in hot o-dichlorobenzene. pDTS1 with 2-ethylhexyl groups showed higher solubility than pDTS2, presumably because of its molecular weight being lower than that of pDTS2 and the branched 2-ethylhexyl chains on the silicon atom. The branched 2-ethylhexyl groups hindered the interchain interaction of the polymers to improve the solubility. pDTS1 was moderately soluble in o-dichlorobenzene, chlorobenzene, THF and toluene, but insoluble in acetone, hexane, ethyl acetate and ethanol. pDTC2 showed even higher solubility; it was soluble in ethyl acetate, hexane, THF, toluene, chlorobenzene and o-dichlorobenzene, but not in ethanol. This was also likely because of its molecular weight being lower than that of pDTS2.

These polymers were characterized mainly by NMR analysis (for the 1H NMR spectra, see Supplementary Figures S1–S3 in Supplementary Information). The 1H and 13C NMR spectra of the polymers were consistent with those based on the regular structures shown in Scheme 2, although sp2 carbon signals could not be clearly observed for the present polymers because of their low solubility. The thermal stability of the present polymers was confirmed by thermogravimetric analysis; the polymers showed 5% weight loss at temperatures (Td5) higher than 350 °C in nitrogen, similar to pDTG1 and pDTG2 (Table 1).15 The Td5 values were much higher than that reported for pDTS314 by more than 100 °C, likely reflecting the higher molecular weights.

Optical and electrochemical properties of the present polymers

The UV–vis absorption spectra of polymers pDTS1, pDTS2 and pDTC2 were measured in chlorobenzene or o-dichlorobenzene solution and as spin-coated films (Figure 1). The UV–vis spectra of the present polymers showed red-shifted absorption bands relative to those of regio-regular P3HT (Aldrich Sigma Inc., St Louis, MO, USA, head-to-tail content >95%, λmax=458 nm). This indicates that the introduction of the carbon and silicon bridging units enhances the conjugation of polythiophenes, similar to the previously reported pDTS314 and DTG polymers.15 Two broad bands were observed at ∼570 and 600 nm when pDTS1 was measured in solution. The band at shorter wavelength was relatively suppressed and a new peak at ∼650 nm appeared when the polymers were measured as films. The new absorption bands in the low energy region reflected the interchain interactions in the aggregated polymer chains in film. The absorption bands in the high energy region were because of the aggregation-free polymer segments. In contrast, the absorption spectrum of pDTS2 in solution showed a shoulder peak at ∼670 nm arising from aggregation already in solution that was enhanced when measured as film, indicating that pDTS2 exhibited stronger interchain interaction than pDTS1. The absorption spectra of pDTC2 showed no significant changes regardless of the state (in solution or film), suggesting the low degree of aggregation in film compared with pDTS2 and pDTG2. The UV absorption maxima of the polymers in solution were shifted to longer wavelengths in the order of pDTG1L<pDTG1H<pDTC2<pDTS1<pDTS2<pDTG2L<pDTG2H. Polymer pDTS2 showed the most red-shifted absorption maximum among the present polymers. The high molecular weight of pDTS2 compared with those of pDTS1 and pDTC2 may be a reason for the red-shift. However, the shifts are rather small as pDTG1 and pDTG2 exhibited only small shifts of the absorption maxima depending on the molecular weights.15 The PL spectra of the polymers are presented in Figure 2: the maxima are at nearly the same wavelengths regardless of substituent for pDTS1 and pDTS2, although pDTC2 appeared at slightly lower energy. The absolute PL quantum efficiencies were determined by using an integration sphere to be <2%.

Figure 1
figure 1

Ultraviolet–visible (UV–vis) absorption spectra of (a) pDTS1 in chlorobenzene, (b) pDTS2 in o-dichlorobenzene, and (c) pDTC2 in chlorobenzene (solid line), and those as films (dashed line).

Full size image
Figure 2
figure 2

Photoluminescence (PL) spectra of dithienosilole (DTS) and cyclopentadithiophene (DTC) polymers in solution excited at the absorption maxima (pDTS1 and pDTC2 in chlorobenzene, and pDTS2 in o-dichlorobenzene).

Full size image

Cyclic voltammograms of the polymer cast films containing 10 equiv of tetrabutylammonium perchlorate (TBAP) on platinum electrodes, which were measured in acetonitrile containing 100 mM TBAP as the supporting electrolyte, showed irreversible anodic peaks (Figure 3). The oxidation onset potentials of the present polymer films were similar to those of pDTG; however, pDTG2 with n-octyl groups showed relatively higher HOMO and LUMO levels than the present polymers (Table 1).15 No obvious changes of cyclic voltammogram onset potential depending on the molecular weight were noted.

Figure 3
figure 3

Cyclic voltammograms (CVs) of polymer cast films.

Full size image

To know about the effects of molecular weight on the polymer properties, we separated pDTS2L from pDTS2 as a fraction that was soluble in dichlorobenzene at 40 °C. As listed in Table 1, the UV absorption and PL maxima in solutions moved to higher energies by 5 and 2 nm, respectively, from those of pDTS2 because of the less extended conjugation. However, these shifts are rather small and effects of the molecular weights do not seem to be significant for pDTS2 and pDTS2L. The UV absorption maximum of the film was also blue-shifted, but to a larger extent (11 nm), indicating that the longer molecular weight led to the better interchain interaction.

When the UV absorption and PL spectra of pDTS1 and pDTS2L with similar molecular weights were compared, no clear tendency of positions of the maxima were seen, indicating that the nature of substituents (linear or branched) exerted no unambiguous influences on the polymer electronic states. However, UV absorption spectrum of pDTS2L in solution still showed a shoulder at ∼670 nm ascribed to the polymer aggregation, similar to that of pDTS2. This indicates that the stronger interchain interaction of pDTS2 than pDTS1 is not only because of the larger molecular weight but also the liner substituents (vide supra). Lack of the evident substituent effects on the polymer conjugation for the present DTS polymers led to a conclusion that the blue-shifted absorption PL bands of pDTS3 from those of pDTS1 and pDTS2 were because of the low molecular weight.

Quantum chemical calculations

To understand the differences in electronic properties among pDTC, pDTS, and pDTG, we carried out crystal orbital (CO) calculations of the polymer models simplified by using methyl groups as the substituents on the bridging atoms (pDTC0, pDTS0 and pDTG0) at the level of B3LYP/6-31G(d) on the Gaussian 09 program (Gaussian, Inc., Wallingford, CT, USA). Optimized geometries of the unit cells, and HOCO and LUCO (highest occupied and lowest unoccupied CO) profiles of the polymers are depicted in Figure 4. The optimized geometries have high planarity and interplane angles <1°. The HOCO and LUCO energy levels and the band gaps derived from the calculations are summarized in Table 2. Both HOCO and LUCO of pDTS0 and pDTG0 were more stable than those of pDTC0. The theoretical band gap of pDTC0 was slightly larger than those of pDTS0 and pDTG0. These suggested the higher stability of pDTS and pDTG toward atmospheric oxidation than pDTC despite the more extended conjugation, although in real systems, the optical band gap of pDTG2 (1.9 eV) was slightly larger than those of pDTC2 (1.8 eV) and pDTS2 (1.7 eV).

Figure 4
figure 4

Optimized geometries (left) and HOCO (middle) and LUCO (right) profiles of pDTC0, pDTS0 and pDTG0. HOCO, highest occupied crystal orbital; LUCO, lowest unoccupied crystal orbital.

Full size image
Table 2 HOCO and LUCO energy levels and band gaps of model polymers, derived from DFT calculations at B3LYP/6–31G(d)
Full size table

Semiconducting properties

OFETs using pDTS and pDTC films as the active layer were fabricated to investigate polymer semiconducting properties. The devices showed clear p-type transistor response (Supplementary Figures S4–S6 in Supplementary Information) in contrast to pDTS3 that was reported to be transistor inactive.14 As shown in Table 3, pDTS2 had higher carrier mobility than pDTS1, likely because of the stronger interchain interaction of pDTS2 in the film than that of pDTS1 as indicated by the UV–vis absorption spectra. This was similar to pDTG1 and pDTG2 and it did not seem unusual that the linear substituents enhance the aggregate-forming property of the polymers, likely because of easy interdigitation of the alkyl chains.15 pDTS2L showed slightly lower mobility than that of pDTS2, likely reflecting both the less enhanced conjugation and interchain interaction because of the lower molecular weight. However, the influence was not significant, as observed for pDTG1H/pDTG2H and pDTG1L/pDTG2L. For polymers with the same n-octyl groups, pDTG2 and pDTS2 exhibited higher carrier mobility than pDTC2. It seemed that the strong interchain interaction of pDTG2 and pDTS2 enhanced carrier transport in the polymer film, likely because of the higher flexibility of substituents on the bridging silicon and germanium, arising from longer Si-C and Ge-C bonds and the larger van deer Waals interaction of these heavier elements. It was also noted that polymer pDTC2 showed a low on/off ratio compared with the other polymers.

Table 3 OFET parameters of DTC, DTS and DTG polymers
Full size table

Conclusions

In summary, we synthesized carbon- and silicon-bridged polythiophenes (pDTC2, pDTS1 and pDTS2). Polymers pDTS1 and pDTS2 had high molecular weights, similar to pDTG1 and pDTG2 reported previously, whereas pDTC2 had a low molecular weight. Attempted preparation of pDTC2 with a higher molecular weight by changing the reaction conditions was unsuccessful. The less flexible side chains of pDTC2, which were because of the shorter distance between the substituent carbon and the bridging element, than that of pDTS resulted in the lower solubility of pDTC2, thus suppressing the polymerization. Interestingly, the present polymers showed similar optical and electrochemical properties to pDTG1 and pDTG2. However, the UV–vis spectra indicated that pDTC2 exhibited less tendency to form polymer aggregates than the other polymers with DTS and DTG units.

All the polymers exhibited p-type transistor properties and pDTS2 had the highest carrier mobility, in contrast to pDTS3 that was previously reported to be nontransistor active.14 It was also noteworthy that pDTC2 had much lower activity in OFET than the other polymers. It is speculated that the suppressed aggregate-forming property of pDTC2 is responsible for the low carrier mobility. The bridging elements, substituents and the molecular weights affected the aggregate formation, thus being important factors to enhance the transistor activity of the polymers. Although the transistor activities of the DTS and DTG homopolymers are lower than those recently reported for other low-band gap donor–acceptor copolymers with DTG or DTS as the donor,21, 22 the results from the present work on simple polymers seem to provide the information of how the nature of bridging group 14 elements affects the properties of bridged polythiophenes, and this would be useful for the molecular design of new functional polythiophene-based materials.

scheme 1

Structures of polythiophene derivatives (P3HT, pDTC, pDTS and pDTG) and heteroatom-bridged 2,2′-bithiophene derivatives (cyclopentadithiophene (DTC), dithienosilole (DTS) and dithienogermole (DTG)).

scheme 2

Synthesis of dithienosilole (DTS) and cyclopentadithiophene (DTC) polymers.